Semantic memory refers to the collective store of concepts, meanings, and names, anatomically linked to the lateral and anterior regions of the temporal lobes in humans. The structure of semantic memory has been theoretically described as a network of concept nodes united by a series of labeled semantic connections (Collins and Loftus 1975). The focus of this proposal is to examine the relative impairment of the semantic memory network in two forms of degenerative dementia, Alzheimer's disease (AD) and Semantic dementia (SD). It has been proposed that the loss of semantic memory in both SD and AD is a degradation of the central semantic storage network (Hodges, Salmon, & Butters 1992; Hodges, Patterson et al. 1992). However, alternative studies of AD have suggested that an inability to consciously retrieve items from semantic memory is the cause of the witnessed deficit (Nebes 1992, Rich et al.2002). The first goal of this proposal is to examine the level of automatic spreading activation produced in the semantic networks of AD and SD patients matched for their overall level of impairment on explicit semantic tasks (Specific Aim 1). To achieve this goal, an implicit threshold oral reading semantic priming task is proposed (Glosser & Friedman 1991). Nakamura et al. (2000) performed a similar comparison of AD and SD and found priming for AD patients only. However, a significantly greater degree of explicit semantic impairment in the SD patients confounded the results. The second goal is to explicitly examine the ability of AD and SD patients to access their semantic networks (Specific Aim 2). This will be accomplished using a series of direct and indirect semantic probe questions assessing attribute knowledge retention (Chertkow et al. 1989, Nebes and Brady 1990, Nebes 1992). The third goal is to examine the hierarchical, bottom-up loss of semantic information in AD and SD (Specific Aim 3). Experiment 1 will use three types of semantic relationships (attributes, category coordinates, and category labels) to test this aim. These experiments are clinically relevant in that the findings could improve differential diagnosis and rehabilitative therapies for AD and SD. Additionally, the results could provide evidence as to the locus of semantic memory processing. [unreadable] [unreadable]